大型工业CT流水线图像重建

Study on pipelining image reconstruction in large industrial CT

大型工业CT中,采用等角度扇形束扫描,为达到高分辨率,需要在每个投影下微动探测器,多次采集数据。为了节省成像时间,设想让重建计算和数据采集同时进行,称为流水线图像重建法。研究了这种流水线图像重建方法,并与传统方法的效率进行了比较。

In the large industrial computerized tomography, we place the detectors in a circle with equal angle between eath other. But the detectors can not be made so small to fit our need, so we have to make them angle little after each collection so as to detect more radial. For economizing time, we project to calculate the partial projections just after they were detected and before all projections were detected. Then our task in this paper is to reconstruct the image using convolution algorithms for divergent beams synchronous with detecting. For this we will first simulate the projections, then reconstruct them group by group, and in the end, compare it with the traditional reconstruction algorithms, now we can find that it is much quicker than them.